Optimization-Based Motion Synthesis for Unified Manipulation in Robot Hand-Arm Systems with Bowden Cable Transmission

Unified hand–arm robot systems face critical challenges in reliable force transmission, particularly when tendon-driven dexterous hands are actuated via Bowden cables. The remote actuation introduces nonlinear friction, hysteresis, and force loss along the transmission path, which compromise coordinated control between the hand and arm. To mitigate these effects, we propose a nonlinear optimization-based motion synthesis framework that minimizes Bowden cable force variation while maintaining precise end-effector tracking. The framework incorporates multiple objectives, including end-effector pose matching, motion smoothness (penalizing velocity, acceleration, and jerk), kinematic singularity avoidance, joint torque minimization, and Bowden cable bending minimization. We evaluated this motion synthesis framework on a single-cable setup as a representative of a multi-tendon, multi-fingered hand-arm system, demonstrating effective reduction of cable force fluctuations and accurate trajectory tracking.

keywords: Task and motion planning, Robotic grasping and manipulation, Mechatronic system modelling, design, optimization